Although nerve conduction study (NCS) is the method most frequently used in daily clinical practice to confirm clinical diagnosis of Carpal tunnel syndrome (CTS), ultrasonographic (US) measurement of the median nerve cross-sectional area is both sensitive and specific for the diagnosis of CTS. Moreover, an algorithm evaluating CTS severity based on CSA of median nerve was suggested. This study is aimed to investigate the clinical usefulness of this algorithm in assessing CTS severity. The patients underwent a full clinical examination, including Tinel and Phalen test, and questioned about symptoms and the secondary causes of CTS. All of the patients refilled a Turkish version Levine Boston Carpal tunnel syndrome questionnaire (BQ) and the visual analog scale for pain (VAS 0-100 mm) A MyLab 70 US system (Esaote Biomedica, Genoa, Italy) equipped with a broadband 6-18 MHz linear transducer was used for US examination. The cross-sectional area of the median nerve was measured at the proximal inlet of the carpal tunnel (US cut-off points that discriminate between different grades of CTS severity as 10.0-13.0 mm(2) for mild symptoms, 13.0-15.0 mm(2) moderate symptoms and >15.0 mm(2) for severe patients). Nerve conduction studies were carried out, and severity of electrophysiological CTS impairment was reported as normal, mild, moderate, severe and extreme. The agreement between NCS and US in showing CTS severity (normal, mild, moderate and severe) was calculated with Cohen's kappa coefficient. Ninety-nine wrists of 54 patients (male/female: 4/50) were included in the study. Mean ages of patients were (+/-SD) 43.3 +/- 11 years. Forty-nine patients had idiopathic CTS, whereas five had secondary CTS (4 had diabetes mellitus and 1 had hypothyroidism). Symptoms were bilateral in 45 patients (83.3%). There were statistical differences between the groups according to electrophysiologic severity scale in terms of age (P < 0.001), body-mass index (P = 0.034), VAS (P = 0.014), Boston symptom severity (P = 0.013) and CSA of median nerve (P < 0.001). The identification of CTS severity showed substantial agreement (Cohen's kappa coefficient = 0.619) between the US and NCS. Also the four groups based on US CTS severity classification were significantly different in VAS (P = 0.017) and Boston symptom severity (P = 0.021). The median nerve swelling detected by calculation of the CSA reflects in itself the degree of nerve damage as expressed by the clinical picture. In addition to CTS diagnosis, sonographic measurement of CSA could also give additional information about severity of median nerve involvement. Using of US may cost-effectively reduce the number of NCS in patients with suspected CTS.